aria-patterns

aria-patterns is a skill for Claude Code, Codex from weAAAre/a11y-agents-kit. It costs 236 tokens per session (2,122 once invoked), scanned A, original, MIT.

A guide to building web interface components that work with screen readers and other assistive technology. It combines ARIA accessibility guidance with information about where real devices and browsers behave differently.

In plain words
What is it for?
Use it when creating dialogs, menus, tooltips, forms, autocomplete fields, and other interactive components. It provides compatibility details and workarounds for known support problems.
Why use it?
It helps prevent interfaces that follow written accessibility rules but still confuse or hide information from users of assistive technology.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/weaaare/a11y-agents-kit/aria-patterns
Any agent
npx skills add weAAAre/a11y-agents-kit --skill aria-patterns
Clone the repo
git clone --depth 1 https://github.com/weAAAre/a11y-agents-kit

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for aria-patterns

README.md
[![agentmods](https://agentmods.dev/badge/skills/weaaare/a11y-agents-kit/aria-patterns.svg)](https://agentmods.dev/skills/weaaare/a11y-agents-kit/aria-patterns)
Your own site
<a href="https://agentmods.dev/skills/weaaare/a11y-agents-kit/aria-patterns"><img src="https://agentmods.dev/badge/skills/weaaare/a11y-agents-kit/aria-patterns.svg" alt="Measured on agentmods" height="20"></a>
Per session 236 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,122 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00236 $0.02122
Opus 5 $0.00118 $0.01061
Sonnet 5 $0.00047 $0.00424
Haiku 4.5 $0.00024 $0.00212

Measured 4d ago against content hash 4cd4f716ae70, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

aria-patterns scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/aria-patterns/SKILL.md · 116 lines

How it starts

The opening of the file, as written. The whole thing — 116 lines — stays where its author put it; the contents beside it link to each section on GitHub.

ARIA Accessible Component Patterns

This skill provides deep, practical knowledge about accessible UI component patterns based on the WAI-ARIA Authoring Practices Guide (APG).

Why this skill exists

LLMs already know the ARIA spec reasonably well. What they often get wrong is the gap between spec and reality:

  • role="tooltip" has 3/33 MUST support on a11ysupport.io — it's essentially dead
  • aria-modal="true" is not supported by TalkBack, Narrator, or Orca — your modal isn't actually modal
  • aria-autocomplete has 11/55 MUST support — screen readers mostly ignore it
  • aria-errormessage has ~20% MUST support — designed for form errors but AT ignores it; use aria-describedby instead
  • aria-relevant is completely broken for ANY non-default value"additions" alone has 0/11 full support (NVDA ZERO on all browsers — stops announcing entirely); "text" alone also 0/11 (Narrator, VoiceOver iOS/macOS = none); "removals" 2/11 (VoiceOver only); the default "additions text" is the ONLY value that works (11/11)
  • aria-controls navigation works only in JAWS — other screen readers don't expose it
  • role="menu" is for application commands, not site navigation — this is the #1 ARIA misuse
  • aria-busy="true" is supposed to suppress live region updates but only JAWS supports it — every other screen reader ignores it
  • aria-live="assertive" does NOT interrupt in JAWS — the most-used desktop screen reader queues it instead
  • aria-details has 7/11 MUST support — VoiceOver (macOS + iOS), Narrator, and TalkBack completely ignore it; use aria-describedby as primary link to descriptions
  • aria-activedescendant is completely unsupported for combobox on VoiceOver macOS/Safari — it works for menus but combobox options are invisible; use DOM focus (roving tabindex) instead
  • aria-current has ZERO support on Narrator and Orca — the attribute recommended for breadcrumbs, step indicators, and nav is invisible to Windows built-in and Linux screen readers; dynamic state changes only announced by NVDA and VoiceOver iOS
  • aria-pressed="mixed" has near-zero support — only JAWS (Chrome/Firefox) and VoiceOver macOS convey the tri-state; NVDA, Narrator, Orca, TalkBack, VoiceOver iOS all ignore it
  • aria-sort dynamic changes are announced by only NVDA+Firefox (1/11 combos) — when a user clicks a column header to change sort direction, JAWS, Narrator, VoiceOver (all platforms), Orca, and TalkBack are all silent; Orca and TalkBack can't even read static ascending/descending values
  • aria-colcount="-1" (unknown column count) is broken in ALL screen readers (0/11) — every AT ignores the spec's MUST NOT and announces its own calculated count, misleading users; VoiceOver (macOS + iOS) and TalkBack don't support aria-colcount at all
  • aria-flowto has 10/44 MUST support (23%) — only JAWS supports alternate reading order navigation; NVDA, VoiceOver, TalkBack, Narrator, and Orca completely ignore it, making it useless for defining multi-column or complex layout reading paths
  • aria-level above 6 has ZERO support in JAWS (all browsers) and TalkBack — only 4/11 combos fully support heading levels >6; JAWS announces "heading" with no level number; restructure headings to stay within 1–6 or encode depth context in the accessible name

Read the full file on GitHub · 116 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 4d ago First seen · 116 lines · 236 tokens per session scan A 4cd4f716ae70

Subscribe to this mod's changes

aria-patterns is a skill published in the GitHub repository weAAAre/a11y-agents-kit (33 stars, last pushed 4mo ago), licensed MIT. It adds 236 tokens to every session and 2,122 once invoked, about $0.0012 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.